What does it take to find the Antikythera Mechanism? Underwater robots, of course!

Meet the chap behind the controls

I worked on ... A few days from now, Christian Lees will be in the Greek islands, sunning himself on the deck of a colossal private yacht. Staff on the yacht will prepare his meals, even do his laundry.

But those staff can't program or troubleshoot Autonomous Underwater Vehicles (AUV), which is why Lees will be aboard the yacht during its second voyage to hunt down fragments of the Antikythera Mechanism.

The Mechanism intrigues boffins from many disciplines, because it was made in about 200 BC and is thought to compute the positions of the planets with a system of meshed gears. Humanity is not known to have constructed a comparable artefact for more than a millennium after its manufacture.

The Mechanism is named for the Greek island of Antikythera, where it was found in the sunken wreck of a Roman ship. In 2014, an expedition visited the site looking for more artefacts and found a fetching statue, but no more ancient clockwork.

This year, Lees tells The Reg, the robots he tends will conduct a magnetic survey of the wreck site. One is a remotely operated vehicle (ROV) operating on a fibre optic tether and controlled from the surface. A second craft, SIRIUS, will conduct a photographic survey of the sea floor.

If either vessel finds signs of bronze on the sea bed, that's where divers will be directed.

The 40 year-old Australian finds himself hunting for the Mechanism after a series of happy career accidents. An electrical engineer by training, he found work at Australia's Defence Science Technology Organisation, where his knowledge of analogue electronics made him suited to work on kit intended for use beneath the waves. Everything's gone digital since, but Lees' career was soon up and swimming in unexpected directions.

After a few years at DSTO he was offered a secondment at The Australian Centre for Field Robotics (ACFR), a facility at the University of Sydney that works on robotics in all manner of industry settings. As is often the case, that secondment turned into a job offer. And now Lees is off to Greece.

The Antikythera gig is a good one because not many of his jaunts are on luxury yachts. The better ones are on research vessels kitted out with all the power and networking needed to make his job easier. The ACFR is the Australian government's robotics outfit of choice for tasks like fisheries surveys, so Lees has spent plenty of time on the Australian coast. His work has also taken him to Japan.

Life aboard ship is busy. Lees spends most of his time working with SIRIUS, the ACFR's main AUV. That could mean dealing with a bent prop shaft, working with watertight seals or routing cabling through the interior. He also works on SIRUIS' PC 104 single-board computers, which he loads with custom Linuxes he's based on Debian and prepared on dry land. C is his preferred language and he wields it to write and integrate drivers as SIRUIS acquires new sensors.

The AUV has a stable operating environment and suite of programs, but things can go awry for any number of reasons. Sometimes AUVs snag. Sometimes their buoyancy becomes an issue. Or new code can be needed between missions. When any of the above strike – and Lees says each trip throws up one big problem – there's all-nighters to be endured and pressure aplenty.

Before each day's mission, SIRIUS is programmed with a path it is hoped it will traverse. But once beneath the waves GPS becomes useless. A suite of other sensors helps the vessel to navigate, but there's only an anaemic 13kbps acoustic modem to tell Lees and his colleagues if the 'bot is on track.

Youtube Video